J Physiol 584.

3 (2007) pp 1019–1028 1019

Inspiratory muscle training attenuates the human

respiratory muscle metaboreflex
Jonathan D. Witt1 , Jordan A. Guenette1 , Jim L. Rupert1 , Donald C. McKenzie1,2 and A. William Sheel1
1
School of Human Kinetics and 2 Faculty of Medicine, The University of British Columbia, Vancouver, British Columbia, Canada

We hypothesized that inspiratory muscle training (IMT) would attenuate the sympathetically
mediated heart rate (HR) and mean arterial pressure (MAP) increases normally observed during
fatiguing inspiratory muscle work. An experimental group (Exp, n = 8) performed IMT 6 days
per week for 5 weeks at 50% of maximal inspiratory pressure (MIP), while a control group (Sham,
n = 8) performed IMT at 10% MIP. Pre- and post-training, subjects underwent a eucapnic
resistive breathing task (RBT) (breathing frequency = 15 breaths min−1 , duty cycle = 0.70)
while HR and MAP were continuously monitored. Following IMT, MIP increased significantly
(P < 0.05) in the Exp group (−125 ± 10 to −146 ± 12 cmH2 O; mean ± S.E.M.) but not in the
Sham group (−141 ± 11 to −148 ± 11 cmH2 O). Prior to IMT, the RBT resulted in significant
increases in HR (Sham: 59 ± 2 to 83 ± 4 beats min−1 ; Exp: 62 ± 3 to 83 ± 4 beats min−1 ) and
MAP (Sham: 88 ± 2 to 106 ± 3 mmHg; Exp: 84 ± 1 to 99 ± 3 mmHg) in both groups relative
to rest. Following IMT, the Sham group observed similar HR and MAP responses to the RBT
while the Exp group failed to increase HR and MAP to the same extent as before (HR: 59 ± 3 to
74 ± 2 beats min−1 ; MAP: 84 ± 1 to 89 ± 2 mmHg). This attenuated cardiovascular response
suggests a blunted sympatho-excitation to resistive inspiratory work. We attribute our findings
to a reduced activity of chemosensitive afferents within the inspiratory muscles and may provide
a mechanism for some of the whole-body exercise endurance improvements associated with IMT.
(Received 17 July 2007; accepted after revision 11 September 2007; first published online 13 September 2007)
Corresponding author A. W. Sheel: Health and Integrative Physiology Laboratory, University of British Columbia,
6108 Thunderbird Blvd, Vancouver, BC, Canada, V6T 1Z3. Email: bill.sheel@ubc.ca

Fatiguing work of the inspiratory muscles is associated
with significant neural and cardiovascular consequences.
Induction of inspiratory muscle fatigue in healthy humans
by means of voluntary resistive inspiration has been
shown to result in time-dependent increases in muscle
sympathetic nerve activity, heart rate (HR) and mean
arterial pressure (MAP) (St Croix et al. 2000), and a gradual
reduction in arterial blood flow to the resting limb (Sheel
et al. 2001). Furthermore, research in the anaesthetized rat
has shown fatiguing contractions of the diaphragm to be
responsible for an increase in the activity of type IV afferent
fibres associated with increased sympathetic outflow (Hill,
2000). Collectively, these findings suggest the existence of
a sympathetically mediated metaboreflex, or chemoreflex,
that originates from fatigued inspiratory muscles. There
is also evidence to suggest that the inspiratory muscle
metaboreflex becomes active with whole-body exercise.
Leg blood flow during high-intensity exercise has been
shown to be inversely related to the work of breathing,
and the consequent changes in leg vascular resistance
have been shown to be directly related to the degree of
noradrenaline (norepinephrine) spillover (Harms et al.
1997). More recently it has been demonstrated that a trans-
ient infusion of lactic acid into the phrenic circulation
reduces limb blood flow and raises MAP in resting
or exercising canines (Rodman et al. 2003). During
prolonged intense whole-body exercise in healthy humans,
inspiratory muscle fatigue can occur (Johnson et al. 1993)
and the associated metaboreflex response may therefore
limit exercise performance under such conditions.
Inspiratory muscle training (IMT) has been associated
with significant improvements in whole-body exercise
performance but the mechanism(s) behind this
phenomenon have not been made clear. Previous
authors have documented increased cycling endurance
(Spengler et al. 1999; Stuessi et al. 2001) whereas others
have not observed any consistent improvement in exercise
performance (Fairbarn et al. 1991; Sonetti et al. 2001;
Guenette et al. 2006). IMT has been associated with
favourable changes in exercising lactate levels (Spengler
et al. 1999), diaphragm thickness (Downey et al. 2006),
and inspiratory muscle fatigue resistance (Boutellier et al.
1992; Boutellier & Piwko, 1992) and strength (Sonetti
et al. 2001). Based on the above brief summary, we asked


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society DOI: 10.1113/jphysiol.2007.140855
1020 J. D. Witt and others
if the cardiovascular effects of the inspiratory muscle
metaboreflex could be attenuated with IMT. Specifically,
we hypothesized that following IMT, subjects would
exhibit an attenuated HR and MAP response during a
resistive breathing task.
Methods
Subjects
All experimental procedures and protocols were approved
by the Clinical Research Ethics Board of the University
of British Columbia and conformed to the Declaration
of Helsinki. Sixteen men provided informed written
consent prior to beginning the study. Subjects were young
(25.8 ± 0.8 years), of normal mass (76.9 ± 2.6 kg) and
height (179.0 ± 2.1 cm) and had a healthy range for body
mass index (23.9 ± 0.6 kg m−2 ). All subjects were free from
cardiovascular, pulmonary and neurological disease. For
screening purposes, all subjects performed pulmonary
function testing using a portable spirometer (Spirolab II,
Medical International Research, Rome, Italy). All subjects
had a normal forced vital capacity (5.4 ± 0.3 l) and forced
expired volume in 1 s (4.5 ± 0.2 l) (Knudson et al. 1983).
General procedures
Mouth pressure (P m ) was measured with a pressure
transducer (Model MP45–36-871, Validyne, Northridge
CA, USA). The end-tidal partial pressure of CO2
(PET,CO2 ) was determined using a gas analyser (CD-3A,
AEI Technologies, Pittsburgh, PA, USA). Inspiratory
flow was determined using a pneumotachograph
(Hans-Rudolph 3813, Kansas City, MO, USA). Each of
these three devices was calibrated prior to every test.
From the flow signal, tidal volume (V T ) and breathing
frequency (f b ) values were determined and minute
ventilation (V̇E ) was calculated as the product of f b and
V T . Offline integration of the P m signal allowed for
the calculation
 of inspiratory muscle force development
(f b × P m ). Arterial blood pressure and HR were
monitored continuously at the finger on a beat-by-beat
basis with the use of a finger arterial plethysmograph
(Finometer, FMS, Finapres Medical Systems BV, Arnhem,
the Netherlands). Systolic and diastolic blood pressures
were measured and MAP was calculated as 1/3 pulse
pressure + diastolic pressure. Blood pressure was also
determined every minute with the use of an auto-
mated blood pressure cuff (BPM-100, VSM MedTech Ltd,
Vancouver, Canada). The average blood pressure values
obtained from the automatic cuff over the resting eupnoeic
period were used to correct the arterial plethysmograph
values for each subject. All signals were sampled at 100 Hz
using an analog-to-digital converter (PowerLab/16SP
J Physiol 584.3
model ML795, ADI, Colorado Springs, CO, USA) and
stored on a computer for subsequent analyses.
Experimental protocol
Subjects were assigned to an experimental (Exp) or a
control (Sham) IMT group. The Exp group trained at
50% of MIP, 6 days a week, for 5 weeks. The Sham sub-
jects performed IMT at 10% of MIP, 6 days a week, for
5 weeks. Subjects were blinded to their assigned treatment
group. Prior to the commencement of the training
program, subjects underwent familiarization testing in
order to minimize any potential learning effect. On the
experimental days (pre- and post-IMT) subjects under-
went a resisted breathing task (see below) designed to
elicit the respiratory muscle metaboreflex (Sheel et al.
2001). MIP and maximal expiratory pressure (MEP) tests
were also performed on these days prior to the resisted
breathing task. Measurements of MEP were made to ensure
the expiratory muscles remained untrained.
Respiratory muscle strength testing
MIP and MEP tests required subjects to perform a maximal
volitional inspiratory or expiratory effort through a
mouthpiece attached to an occluded three-way stopcock
(2100 series, Hans Rudolph, Kansas City, MO, USA).
A small pinhole within the stopcock served to prevent
glottic closure and minimize the recruitement of the buccal
muscles. The MIP and MEP tests were performed in a
seated position and conformed to standarized procedures
(ATS, 2002). Inspiratory manoeuvers were initiated from
residual volume and expiratory manoeuvers were initiated
from total lung capacity. Three acceptable MIP and MEP
values were each averaged and the resulting means were
taken as the test values.
Resistive breathing task
Subjects refrained from caffeine, exercise, alcohol and
food for 12 h prior to each RBT. All RBTs took place
with subjects in a semi-recumbant position. Each trial
began with 8 min of eupnoea in order to ensure stable
baseline cardiorespiratory parameters. The RBT trials
were designed to increase the metabolic requirements
and compromise diaphragm perfusion by combining
a prolonged breath duty cycle and high force output.
Subjects inspired (custom-made Y valve) against an added
resistive load to a target mouth pressure of 60% MIP
by following a tracing of P m on a computer monitor
to a pre-set target for each inspiration. Throughout
the trial, subjects maintained a f b of 15 breaths min−1
and a prolonged duty cycle (T I /T TOT = 0.7) by listening
to a computer-generated audio signal with distinct

C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
J Physiol 584.3 Cardiovascular effects of training respiratory muscle
inspiratory and expiratory tones. This protocol has
previously been shown to reduce diaphragmatic blood
flow in anaesthetized dogs (Bellemare et al. 1983) and
predict the onset of diaphragm fatigue in humans
(Bellemare & Grassino, 1982a,b; Sheel et al. 2001, 2002).
The maximal P m of each breath was continously displayed
for the subjects on a computer monitor. Subjects were
instructed to isolate the diaphragm during inspiration
and minimize accessory breathing muscle recruitment.
Subjects were verbally encouraged to maintain proper
timing, breathing technique and P m as needed.
Termination of the RBTs was based upon the presence
of a plateau in the rise in MAP. The RBTpost was
performed at the same absolute intensity and duration
as the RBTpre . Each RBT was followed by a minute of
non-resistive recovery. Inspired CO2 was increased as
needed to maintain PET,CO2 at eucapnoeic levels.
Inspiratory muscle training
Following the pre-test, subjects were assigned a respiratory
muscle trainer (Powerlung ‘Sport’, Vacumed, Ventura, CA,
USA) set at a resistance corresponding to 50% (Exp) or
10% (Sham) of their seated baseline MIP. Subjects in the
Exp group were encouraged to inspire briskly whereas
subjects in the Sham group inspired normally. To ensure
the subjects did not experience expiratory resistance,
subjects removed the device during expiration. In
addition, subjects were encouraged to practice slow and
relaxed expirations, each lasting 4–6 s. Subjects were
instructed to perform IMT sessions once a day, 6 days
a week, for 5 weeks. Each IMT session involved three
sets of 75 breaths with a 5 min rest between sets. The
training protocol used by the Exp group in this study
was based closely upon an IMT regimen previously shown
to elicit significant improvements in MIP (Sonetti et al.
2001). Subjects reported to the laboratory for supervised
training sessions once a week, during which time seated
MIP and MEP values were re-assessed, and the resistance
on the trainers was adjusted to maintain the same relative
workload. Subjects kept daily logs of their IMT activity
indicating the date and time of their training sessions.
Data and statistical analysis
Spirometry values, descriptive characteristics, and baseline
MIP and MEP values were compared between the Sham
and Exp groups using independent t tests. A dependent  however, MEP values were significantly greater (P < 0.01)
t test was used to assess for possible differences in f b × P m
between RBTpre and RBTpost . MIP and MEP values for each
group were compared from pre- to post-training using
one-tailed paired t tests. For the analysis of the RBT data,
mean values were calculated for each minute of the RBTs
of each subject. A single mean value was calculated for

C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
1021
both the entire 8 min resting period and the first minute
of recovery. Due to variation in the duration of RBTs
between subjects, select points (eupnoea, minutes 1–3, the
final minute and first minute of recovery) were used for
statistical comparisons. For each RBT, repeated measures
ANOVA were performed to detect significant changes from
eupnoeic values in those variables measured. Tukey’s post
hoc test was used when differences were discovered. To
compare the HR and MAP data pre- to post- for each
time point in each group, one-tailed dependent t tests
with Bonferroni corrections were performed. An alpha
level of 0.05 was used for all tests of significance. All
data are expressed as means ± s.e.m. All statistics were
performed with the use of commercial statistical software
(STATISTICA, Version 6.1, Statsoft Inc., Tulsa, OK,
USA).
Results
Baseline measurements and subject compliance
Subjects in the Sham and Exp groups were similar for lung
function measures and physical characteristics (P > 0.05).
During eupnoeic breathing there were no systematic trends
for any measured respiratory variables (Tables 1 and 2,
P > 0.05). The within-subject coefficients of variation
for resting HR and MAP were in excellent agreement
with previously reported values (Sheel et al. 2001). The
average resting coefficients of variation for HR over the
2 days was ±4.1% and ±2.4% for the Sham and Exp
groups, respectively. The resting coefficients of variation
for MAP were similarly low for the Sham (±2.2%) and
Exp (±1.9%) groups. Based on the self-report training
logs, subjects in the Sham group maintained a mean
IMT compliance of 88% over the course of the 5 weeks.
Exp subjects reported 97% compliance; however, this was
not statistically different from that of the Sham subjects
(P > 0.05). The trend for a lower training compliance in
the Sham group was driven primarily by one subject who
exhibited a compliance of only 40%. With data from this
subject removed, the training compliance of the Sham
group was greater than 95%.
Respiratory muscle strength
Baseline MIP strength was similar between the Exp
and Sham groups prior to training (P > 0.05) (Exp:
−125.0 ± 10.0 cmH2 O; Sham: −141.2 ± 11.2 cmH2 O);
in the Sham group (Sham: 158.9 ± 7.4 cmH2 O; Exp:
127.2 ± 6.2 cmH2 O). The MIP and MEP training
responses for both groups are displayed in Fig. 1.
A significant increase in MIP following training was
observed in the Exp group but not the Sham group. All
subjects in the Exp group displayed some rise in MIP after
1022 J. D. Witt and others J Physiol 584.3

Table 1. Group mean data for the Sham group during resistive breathing
fb
VT (breaths V̇E Peak Pm VT /TI PET,CO2 SBP DBP
(l) min−1 ) (l min−1 ) T I /T TOT (cmH2 O) (l s−1 ) (mmHg) (mmHg) (mmHg)

Pre
Eupnoea 0.768 14.10 9.2 0.343 −0.8 0.54 37.3 117.5 73.4
±0.074 ±1.17 ±0.6 ±0.014 ±0.2 ±0.05 ±0.9 ±1.8 ±2.1
Min1 1.190∗ 14.97 16.0∗ 0.736∗ −63.2∗ 0.41 33.8∗ 122.9 74.4
±0.069 ±0.03 ±0.9 ±0.005 ±6.5 ±0.02 ±0.6 ±5.1 ±2.9
Min 2 1.163∗ 15.01 15.7∗ 0.739∗ −62.0∗ 0.39 38.1 136.0∗ 85.4∗
±0.062 ±0.01 ±0.8 ±0.005 ±4.5 ±0.02 ±0.6 ±4.2 ±2.6
Min 3 1.129∗ 15.24 15.4∗ 0.736∗ −60.3∗ 0.39 38.4 134.7∗ 84.6∗
±0.059 ±0.16 0.9 ±0.008 ±5.4 ±0.02 ±1.0 ±4.9 ±3.0
End 1.157∗ 14.98 15.6∗ 0.738∗ −68.6∗ 0.39 38.0 141.7∗ 88.4∗
±0.076 ±0.02 ±1.0 ±0.008 ±6.2 ±0.02 ±0.8 ±4.6 ±2.1
Rec 1 1.164∗ 16.67 16.0∗ 0.366 −0.8 0.93∗ 37.3 129.8∗ 80.5
±0.122 ±1.55 ±1.3 ±0.023 ±0.1 ±0.12 ±1.9 ±3.2 ±2.7

Post
Eupnoea 0.994 10.67 9.1 0.328 −0.6 0.58 35.0 117.7 75.0
±0.134 ±0.57 ±1.2 ±0.022 ±0.1 ±0.09 ±1.8 ±4.2 ±2.9
Min 1 1.157 14.92∗ 15.6∗ 0.726∗ −68.3∗ 0.40 32.7 120.7 74.3
±0.081 ±0.03 ±1.1 ±0.008 ±6.8 ±0.03 ±1.6 ±5.0 ±2.5
Min 2 1.115 15.03∗ 15.1∗ 0.737∗ −66.8∗ 0.38 35.0 132.2∗ 84.5∗
±0.078 ±0.01 ±1.1 ±0.007 ±6.0 ±0.03 ±1.8 ±4.8 ±2.7
Min 3 1.093 15.00∗ 14.8∗ 0.721∗ −65.2∗ 0.38 36.6 132.7∗ 85.5∗
±0.071 ±0.01 ±1.0 ±0.012 ±5.6 ±0.02 ±1.6 ±4.7 ±2.9
End 1.080 15.05∗ 14.7∗ 0.696∗ −71.6∗ 0.39 36.2 135.4∗ 86.1∗
±0.064 ±0.02 ±0.9 ±0.012 ±6.9 ±0.02 ±1.7 ±4.9 ±2.4
Rec 1 1.097 16.76∗ 16.0∗ 0.364 −0.8 0.93∗ 34.7 125.7 80.6
±0.086 ±1.39 ±1.6 ±0.022 ±0.1 ±0.13 ±1.8 ±3.2 ±3.2
Values are means ± S.E.M. (n = 8). ∗ Significantly different from eupnoea (P < 0.05). Definition of terms: V T , tidal volume; f b , breathing
frequency; V̇E , minute ventilation; T I /T TOT duty cycle [inspiratory time (T I )/total time (T TOT )]; Peak P m , peak mouth pressure; V T /T I ,
mean inspiratory flow rate; PET,CO2 , end-tidal carbon dioxide; SBP, systolic blood pressure; DBP, diastolic blood pressure; Min1–3,
minutes 1–3 of resistive breathing; Rec1, first minute of recovery.

5 weeks of training, with a group mean improvement of
17%. Although there was a trend for improvements in
MIP in the Sham group (+6%), this finding was neither
consistent across all subjects, nor significant (P > 0.05).
No significant changes in MEP were observed in either
group following training.
Respiratory measures during resistive breathing
The mean RBT was 535 ± 52 s in duration (range:
266–996 s). Over the course of the RBT, subjects in both
groups had difficulty achieving the target 60% MIP level
with mean peak P m values of 47.3 ± 1.5% and 43.9 ± 3.3%
of MIP in the Sham and Exp groups, respectively. The
RBT duration and P m values were held constant for
pre- and post-IMT trials. These pressure values combined
with the imposed breathing pattern corresponded to a

substantial amount of force with mean f b × P m
values over 200 times eupnoeic levels for both groups
(Fig. 3). All subjects were able to maintain the desired
breathing pattern during the RBTs (f b = 15 breaths min−1 ;
T I /T TOT = 0.70) (Tables 1 and 2). As expected, the
prescribed breathing protocol did result in some degree
of hyperventilation for all subjects during all RBTs, with
f b and V̇E values significantly above rest. This hyper-
ventilation resulted in significant drops in PET,CO2 during
the first 30–60 s of the RBTs in some subjects but was
adequately compensated for with the administration of
supplemental CO2 during subsequent minutes.
Cardiovascular measures during resistive breathing
Prior to training, the resistive breathing resulted in
significant and sustained HR (Sham: +41%; Exp: +35%)


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
J Physiol 584.3 Cardiovascular effects of training respiratory muscle 1023

Table 2. Group mean data for the Exp group during resistive breathing
fb
VT (breaths V̇E Peak P m V T /T I PET,CO2 SBP DBP
(l) min−1 ) (l min−1 ) T I /T TOT (cmH2 O) (l s−1 ) (mmHg) (mmHg) (mmHg)

Pre
Eupnoea 0.781 12.93 8.0 0.304 −0.8 0.54 39.6 113.6 69.9
±0.104 ±1.65 ±0.5 ±0.018 ±0.1 ±0.05 ±1.1 ±1.7 ±1.5
Min 1 1.095∗ 14.99 14.9∗ 0.723∗ −56.8∗ 0.38∗ 35.1∗ 110.7 66.6
±0.058 ±0.03 ±0.8 ±0.013 ±6.8 ±0.02 ±1.0 ±3.8 ±2.9
Min 2 1.068∗ 14.99 14.5∗ 0.722∗ −55.6∗ 0.37∗ 36.6 121.6 74.4
±0.057 ±0.03 ±0.8 ±0.016 ±5.9 ±0.02 ±1.5 ±2.5 ±2.5
Min 3 1.055 15.14 14.4∗ 0.713∗ −59.4∗ 0.37∗ 38.5 126.9∗ 77.1∗
±0.063 ±0.12 ±0.9 ±0.017 ±6.2 ±0.02 ±0.9 ±4.5 ±4.5
End 1.007 14.79 13.6∗ 0.722∗ −51.0∗ 0.35∗ 37.7 133.0∗ 82.3∗
±0.066 ±0.13 1.0 ±0.015 ±5.7 ±0.02 ±0.8 ±3.2 ±2.8
Rec 1 0.942 17.35∗ 13.7∗ 0.348 −0.9 0.80∗ 38.3 124.5∗ 75.9
±0.130 ±1.62 ±1.1 ±0.010 ±0.1 ±0.07 ±2.2 ±2.5 ±1.8

Post
Eupnoea 0.940 10.98 8.5 0.304 −0.5 0.56 36.8 112.8 70.3
±0.117 ±1.04 ±0.9 ±0.014 ±0.1 ±0.05 ±1.6 ±2.1 ±1.2
Min 1 1.079 14.87 14.4∗ 0.689∗ −61.8∗ 0.39∗ 31.6∗ 108.9 63.7∗
±0.023 ±0.12 ±0.3 ±0.018 ±5.2 ±0.01 ±1.4 ±4.0 ±2.8
Min 2 1.080 15.02 14.6∗ 0.709∗ −64.6∗ 0.38∗ 34.9 117.5 70.6
±0.033 ±0.02 ±0.4 ±0.014 ±6.3 ±0.01 ±1.6 ±3.0 ±1.9
Min 3 1.015 15.00 13.7∗ 0.690∗ −63.4∗ 0.37∗ 36.9 121.3∗ 73.6
±0.038 ±0.03 ±0.5 ±0.019 ±7.2 ±0.01 ±1.3 ±3.0 ±2.2
End 0.985 15.01 13.3∗ 0.693∗ −56.4∗ 0.36∗ 36.9 119.7∗ 73.2
±0.049 ±0.04 ±0.6 ±0.014 ±6.3 ±0.02 ±1.8 ±3.1 ±2.1
Rec 1 1.080 16.00∗ 13.2∗ 0.343 −0.8 0.68 34.2 117.7 72.7
±0.178 ±2.24 ±1.5 ±0.016 ±0.1 ±0.07 ±1.7 ±2.1 ±1.2
Values are means ± S.E.M. (n = 8). ∗ Significantly different from eupnoea (P < 0.05). Definition of terms: V T , tidal volume; f b , breathing
frequency; V̇E , minute ventilation; T I /T TOT , duty cycle [inspiratory time (T I )/total time (T TOT )]; Peak P m , peak mouth pressure; V T /T I ,
mean inspiratory flow rate; PET,CO2 , end-tidal carbon dioxide; SBP, systolic blood pressure; DBP, diastolic blood pressure; Min1–3,
minutes 1–3 of resistive breathing; Rec1, first minute of recovery.

Sham
and MAP (Sham: +21%; Exp: +17%) increases in both Exp
training groups. The rise in HR occurred within the
first minute of resistive breathing and tended to return 20 MIP MEP
to near baseline levels within a minute of recovery. The
rises in MAP were more gradual, typically occurring 10
Pressure (cm H2O)

within 2 or 3 min of resistive breathing, and failing

to normalize within the first minute of recovery. 0
Fig. 2A and B show a trace of one representative Exp
subject during the RBTpre and RBTpost and demonstrate -10
an attenuated blood pressure response. The mean HR
and MAP responses for both the Sham and Exp groups -20
are displayed in Fig. 3. Following training in the Sham
group, HR and MAP were again elevated in response -30 *
to the RBT (HR: +36%; MAP: +15%). There were no
differences in HR or MAP between the pre- and post-RBT Figure 1. Changes in maximal inspiratory pressure (MIP) and
maximal expiratory pressure (MEP) following 5 weeks of
tests of the Sham group at any time point (P > 0.05). In inspiratory muscle training in the Sham (n = 8) and Exp (n = 8)
contrast, the Exp group exhibited significantly blunted groups
HR and MAP responses during the RBTpost (HR: +27%; Values are means ± S.E.M. ∗ Post-training value significantly different
MAP: +4%). from pre-training value (P < 0.05).


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
1024 J. D. Witt and others J Physiol 584.3

Discussion
Main findings
The purpose of this study was to investigate the effects
of IMT on the cardiovascular responses to resistive
inspiratory muscle work. To this end, subjects were
required to perform a resistive breathing task before
and after a 5 week period of inspiratory muscle training.
During the RBTpre , subjects in both groups exhibited
a respiratory muscle metaboreflex as evidenced by a
time-dependent rise in HR and MAP. Following training
in the experimental group, the HR response to the same
absolute workload was blunted and the MAP response
was nearly abolished; however, in the Sham group there
were no significant differences in the rise of HR or MAP
between baseline and post-training tests. We interpret this
to indicate a training-induced attenuation of the activity
of chemosensitive afferent fibres in response to resistive
inspiratory muscle work.

Inspiratory muscle strength changes with training

We detected an average MIP improvement of
approximately 21 cmH2 O in the Exp group. This
17% increase in MIP is within the range of mean training
improvements (8–41%) observed by others employing a
similar 50% MIP training protocol (Sonetti et al. 2001;
Guenette et al. 2006; McConnell & Lomax, 2006). The
Sham group results are similar to those seen in other
studies (Sonetti et al. 2001; Volianitis et al. 2001; Downey
et al. 2006). The fact that there were no significant changes
in MEP in either group further suggests that the MIP
improvements observed in the Exp subjects were a result
of the IMT training.

Sympathetic attenuation with inspiratory

muscle training
During the RBTpre it is believed that there were both
central and peripheral factors working to increase HR and
MAP above resting values. The increased central command
associated with the increased inspiratory efforts probably
resulted in significant vagal withdrawal (Hollander &
Bouman, 1975). Additionally, the greater contractile
force associated with the RBT presumably increased the
mechanical deformation of the diaphragm and increased
the activity of the mechanically sensitive (type III) afferent

Figure 2. Raw data from an individual Exp subject during

resistive breathing
Resistive breathing at 60% of MIP, prolonged duty cycle of 0.70 and a
breathing frequency of 15 breaths min−1 at baseline (A) and following
five weeks
 of inspiratory muscle training (B). V T , tidal volume;
f b × P m , inspiratory muscle force generation.


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
J Physiol 584.3 Cardiovascular effects of training respiratory muscle 1025

fibres within this muscle (Duron, 1981; Jammes &
Speck, 1995). Typically, the changes to HR and MAP
induced by mechanoreceptors appear earlier and are of
a lesser magnitude than those of the more delayed-acting
metaboreceptors. The initial HR and MAP responses in
our study are similar to those of others (St Croix et al. 2000;
Sheel et al. 2001) who have manipulated mechanoreceptor
or central command input by employing extremely high
resistive (95% MIP) breathing or hyperpnoea tasks in
the absence of significant metaboreceptor activity. We
therefore believe the HR and MAP increases observed in
the early minutes of the RBT to be mechanoreceptor and
centrally mediated.
The prolonged duty cycle and high inspiratory
resistance of this task has been shown to bring about fatigue
by restricting blood flow to the diaphragm and reducing
its mechanical efficiency (Bellemare & Grassino, 1982a,b;
Bellemare et al. 1983). Over time, an accumulation of lactic
acid and other metabolic by-products is believed to result
in the sympathetically mediated metaboreflex (Dempsey
et al. 2002). The delayed and gradual time-course of the
rise in MAP observed in the Exp subjects prior to training
and in the Sham subjects, both before and after IMT, is
consistent with this reflex as observed by others (St Croix
et al. 2000; Sheel et al. 2001; McConnell & Lomax, 2006).
In contrast, HR failed to respond in a delayed fashion and
observed a more abrupt increase. As HR is more vulnerable
to central influences (Victor et al. 1989), it is possible this
mechanism played a greater role in the HR time-course
and, as a result, masked the role played by the metaboreflex.
We exposed subjects to nearly identical P m , f b , V T and
T I /T TOT values during the pre- and post-RBTs. As a result,
the absolute amount of inspiratory work and the degree
of mechanical deformation experienced by the diaphragm
during the RBTpost was no less than that during the RBTpre .
Despite these control measures, we observed a blunted HR
and MAP response during the RBTpost of the Exp group
in comparison to pre-training values and values in the
Sham group. There are numerous possibilities as to why
this may have occurred. Given the MIP improvements
with IMT, it may be suggested that subjects in the Exp
group were working at a lower relative intensity during the
RBTpost and that this contributed to our findings. However,
this is unlikely to be the case as the MIP improvements
were offset by the slightly greater P m generated during
the RBTpost (Exp: 44.3 ± 10.0% MIP; Sham: 50.2 ± 9.0%

Pre
Post
Sham Exp

0 0
fb x Pm (cm H2O)

fb x Pm (cm H2O)

-10 -10

-20 -20

-30 -30

-40 -40

110 110

*
100 100
MAP (mm Hg)

MAP (mm Hg)

90 90

80 80

70 70

Figure 3. Group mean heart rate, mean 90 90

*
arterial pressure (MAP) and inspiratory
 *
muscle force generation (f b × P m ) during
HR (beats min )

HR (beats min )

80 80
-1

-1

resistive breathing before (•) and after ( e)

inspiratory muscle training in the Sham
70 70
(n = 8) and Exp (n = 8) groups
Eup, eupnoea; Min1–3, minutes 1–3 of
resistive breathing; End, final minute of resistive 60 60
breathing; Rec1, first minute of recovery.
Values are means ± S.E.M. ∗ Significantly 50 50
different from post-training value (P < 0.05). Eup Min1 Min2 Min3 End Rec1 Eup Min1 Min2 Min3 End Rec1


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
1026 J. D. Witt and others J Physiol 584.3

MIP). A second possibility is that there may have been muscles. However, based on studies of isolated arteriole
less mechanoreceptor activity during the RBTpost due to a preparations, it has been suggested that in comparison with
reduced responsiveness to a given mechanical stimulus. the arterioles of the leg, those of the diaphragm exhibit less
This may have occurred due to conditioning of the of a vasoconstrictive response to sympathetic input (Aaker
type III receptors to the repeated mechanical deformations & Laughlin, 2002). This implies that during activation of
experienced during the IMT. Such an adaptation is the metaboreflex there was a relative increase in blood
hypothesized to have occurred in training studies of other flow to the diaphragm in comparison with the locomotor
skeletal muscles (Sinoway et al. 1996; Fisher & White, muscles.
1999). Sinoway et al. (1996) have suggested that this effect
may be related to the muscle acidosis associated with
training, as it has been shown in the anaesthetized feline Relevance for exercise performance
model that the mechanically sensitive type III afferents
Is respiratory muscle work an important determinant of
have a reduced discharge frequency following repeated
the ability to perform dynamic exercise? Two lines of
lactic acid exposure (Sinoway et al. 1993). However,
evidence suggest that activation of the respiratory muscle
mechanoreceptor attenuation if present, should have been
metaboreflex contributes to increased sympathetic tone
apparent from the onset of the RBTpost and would not solely
and redistribution of blood flow during dynamic exercise.
explain the growing disparity between the MAP values of
Firstly, increasing or decreasing the work of breathing
the RBTpre and RBTpost over time.
during heavy cycle exercise has effects on leg blood flow
Although mechanoreceptor adaptations may have
(Harms et al. 1997). It appears there is a threshold or
contributed to the blunted MAP and HR we do not feel that
critical point for activation of this reflex that does not
they entirely explain the attenuated responses observed
occur during submaximal exercise (Wetter et al. 1999) or
during the latter stages of the RBTpost . Alternatively, we
non-fatiguing inspiratory efforts (Sheel et al. 2002). In the
propose that following IMT, the aerobic capacity of the
present study, we believe that IMT may have raised the
respiratory muscles may have been improved thereby
so-called ‘ischaemic threshold’ required to elicit significant
reducing the imbalance between diaphragm blood flow
increases in MAP. Such a threshold change has previously
and metabolic demand during the RBTpost . The resulting
been demonstrated (Mostoufi-Moab et al. 1998) to occur
reduction in metabolite accumulation would attenuate the
at the forearm following a period of handgrip training.
activity of not only the type IV chemosensitive afferents
Secondly, directly altering the work of breathing during
but also the mechanically sensitive type III afferents,
maximal exercise has been shown to have effects on
which evidence suggests are influenced by local metabolite
locomotor muscle fatigue (Romer et al. 2006) and exercise
concentrations as well (Sinoway et al. 1993). Changes to
performance (Harms et al. 2000) in endurance athletes.
the enzyme profile and fibre type composition of the
Indeed, respiratory muscle unloading and loading results
diaphragm have been documented following treadmill
in a 14% improvement and 15% decline, respectively, in
training in rats (Vrabas et al. 1999) and in clinical
the time to fatigue during maximal exercise (Harms et al.
populations that experience chronically high inspiratory
2000). This has also been demonstrated in exercising heart
muscle workloads (Tikunov et al. 1997; Ribera et al.
failure patients (Mancini et al. 1997; O’Donnell et al. 1999).
2003). We believe that such aerobic adaptations of the
Most recently, McConnell & Lomax (2006) have shown
diaphragm were also present following IMT in our Exp
that inspiratory muscle work performed immediately
group. Additionally, during IMT the diaphragm probably
prior to plantar exercise will expedite plantar flexor fatigue
experienced repeated metabolite exposures that may have
and that IMT can raise the threshold of inspiratory
led to a decreased responsiveness of the chemoreceptors
muscle work necessary to elicit this effect. Collectively, our
to a given metabolic stimulus. Such a training effect has
findings coupled with those of others point to important
previously been proposed by Sinoway et al. (1992) who
reflexes originating in the respiratory muscles which have
showed that the forearms of trained individuals exhibited
consequences for exercise performance.
a lesser sympathetic response to a given pH change when
compared with those of untrained individuals. Similar
adaptations at the level of the metaboreceptors in this study
Methodological considerations
would explain the blunted HR and MAP responses during
the final minutes of the RBTpost . There are a number of limitations to this study
Debate exists as to whether or not the respiratory that are worthy of mention. Firstly, we employed a
metaboreflex serves to improve the perfusion of the volitional pressure generation test (MIP and MEP) rather
ischaemic diaphragm. Although the increased sympathetic than an electrically evoked test for the assessment of
stimulus associated with this reflex has been shown inspiratory and expiratory muscle strength. Although
to restrict limb blood flow, the sympathetic outflow is the improvement in MIP in the Exp group strongly
likely to also target the vasculature of the respiratory suggests an increase in global inspiratory muscle


C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
J Physiol 584.3 Cardiovascular effects of training respiratory muscle
strength, it is unclear what physiological adaptations were
behind such increases. Increased central drive, increased
motor unit recruitment, increased inspiratory muscle
contractility, and inspiratory muscle hypertrophy are all
possible explanations for the MIP increase observed after
IMT.
Inspiratory muscle fatigue was not assessed in this study.
Therefore, we cannot be certain as to whether fatigue
did or did not occur. However, based upon the current
understanding of the conditions necessary to induce
inspiratory muscle fatigue (Bellemare & Grassino, 1982a,b;
Bellemare et al. 1983), and based upon the findings
of others who have assessed inspiratory muscle fatigue
following the same RBT protocol (Sheel et al. 2001) as
that employed in this study, we would speculate that the
peak P m values, T I /T TOT values and duration of the RBTs
were sufficient to induce inspiratory muscle fatigue in our
subjects prior to training.
Summary
In summary, our findings demonstrate that 5 weeks of
resistive IMT is capable of increasing inspiratory muscle
strength and attenuating the time-dependent rise in HR
and MAP that occurs with resistive inspiratory work in
healthy males. We attribute the attenuated cardiovascular
response to a reduced activity of chemically sensitive
afferent fibres innervating the inspiratory muscles. Such
an effect may have been caused by an increased oxidative
capacity of the inspiratory muscles, resulting in a
reduced metabolite accumulation with resistive work.
Alternatively, the observed attenuation may be due to
a desensitization or decline in responsiveness of the
type III and IV afferents to chemical stimulants, resulting
from a conditioned response to repeated exposure to
the accumulated metabolites associated with IMT. Our
findings of reduced cardiovascular responsiveness to high
levels of inspiratory work suggest that IMT may have the
potential to improve whole-body exercise performance.
References
Aaker A & Laughlin MH (2002). Diaphragm arterioles are less
responsive to α1-adrenergic constriction than gastrocnemius
arterioles. J Appl Physiol 92, 1808–1816.
ATS (2002). ATS/ERS Statement on respiratory muscle testing.
Am J Respir Crit Care Med 166, 518–624.
Bellemare F & Grassino A (1982a). Effect of pressure and
timing of contraction on human diaphragm fatigue. J Appl
Physiol 53, 1190–1195.
Bellemare F & Grassino A (1982b). Evaluation of human
diaphragm fatigue. J Appl Physiol 53, 1196–1206.
Bellemare F, Wight D, Lavigne CM & Grassino A (1983). Effect
of tension and timing of contraction on the blood flow of the
diaphragm. J Appl Physiol 54, 1597–1606.

C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society
1027
Boutellier U, Buchel R, Kundert A & Spengler C (1992). The
respiratory system as an exercise limiting factor in normal
trained subjects. Eur J Appl Physiol Occup Physiol 65,
347–353.
Boutellier U & Piwko P (1992). The respiratory system as an
exercise limiting factor in normal sedentary subjects.
Eur J Appl Physiol Occup Physiol 64, 145–152.
Dempsey JA, Sheel AW, St Croix CM & Morgan BJ (2002).
Respiratory influences on sympathetic vasomotor outflow in
humans. Respir Physiol Neurobiol 130, 3–20.
Downey AE, Chenoweth LM, Townsend DK, Ranum JD,
Ferguson CS & Harms CA (2006). Effects of inspiratory
muscle training on exercise responses in normoxia and
hypoxia. Respir Physiol Neurobiol 156, 137–146.
Duron B (1981). Intercostal and diaphragmatic muscle
afferents. In Regulation of Breathing, Part I, ed. Hornbein TF,
pp. 473–540. Marcel Decker, New York.
Fairbarn MS, Coutts KC, Pardy RL & McKenzie DC (1991).
Improved respiratory muscle endurance of highly trained
cyclists and the effects on maximal exercise performance.
Int J Sports Med 12, 66–70.
Fisher WJ & White MJ (1999). Training-induced adaptations in
the central command and peripheral reflex components of
the pressor response to isometric exercise of the human
triceps surae. J Physiol 520, 621–628.
Guenette JA, Martens AM, Lee AL, Tyler GD, Richards JC,
Foster GE, Warburton DE & Sheel AW (2006). Variable
effects of respiratory muscle training on cycle exercise
performance in men and women. Appl Physiol Nutr Metab
31, 159–166.
Harms CA, Babcock MA, McClaran SR, Pegelow DF, Nickele
GA, Nelson WB & Dempsey JA (1997). Respiratory muscle
work compromises leg blood flow during maximal exercise.
J Appl Physiol 82, 1573–1583.
Harms CA, Wetter TJ, St Croix CM, Pegelow DF & Dempsey JA
(2000). Effects of respiratory muscle work on exercise
performance. J Appl Physiol 89, 131–138.
Hill JM (2000). Discharge of group IV phrenic afferent fibers
increases during diaphragmatic fatigue. Brain Res 856,
240–244.
Hollander AP & Bouman LN (1975). Cardiac acceleration in
man elicited by a muscle-heart reflex. J Appl Physiol 38,
272–278.
Jammes Y & Speck DF (1995). Respiratory control by
respiratory and diaphragmatic muscle afferents. In
Regulation of Breathing, ed. Dempsey JA & Pack AI,
pp. 543–582. Marcel Dekker, New York.
Johnson BD, Babcock MA, Suman OE & Dempsey JA (1993).
Exercise-induced diaphragmatic fatigue in healthy humans.
J Physiol 460, 385–405.
Knudson RJ, Lebowitz MD, Holberg CJ & Burrows B (1983).
Changes in the normal maximal expiratory flow-volume
curve with growth and aging. Am Rev Respir Dis 127,
725–734.
McConnell AK & Lomax M (2006). The influence of inspiratory
muscle work history and specific inspiratory muscle training
upon human limb muscle fatigue. J Physiol 577, 445–457.
Mancini D, Donchez L & Levine S (1997). Acute unloading of
the work of breathing extends exercise duration in patients
with heart failure. J Am Coll Cardiol 29, 590–596.
1028 J. D. Witt and others
Mostoufi-Moab S, Widmaier EJ, Cornett JA, Gray K & Sinoway
LI (1998). Forearm training reduces the exercise pressor
reflex during ischemic rhythmic handgrip. J Appl Physiol 84,
277–283.
O’Donnell DE, D’Arsigny C, Raj S, Abdollah H & Webb KA
(1999). Ventilatory assistance improves exercise endurance
in stable congestive heart failure. Am J Respir Crit Care Med
160, 1804–1811.
Ribera F, N’Guessan B, Zoll J, Fortin D, Serrurier B, Mettauer
B, Bigard X, Ventura-Clapier R & Lampert E (2003).
Mitochondrial electron transport chain function is enhanced
in inspiratory muscles of patients with chronic obstructive
pulmonary disease. Am J Respir Crit Care Med 167, 873–879.
Rodman JR, Henderson KS, Smith CA & Dempsey JA (2003).
Cardiovascular effects of the respiratory muscle
metaboreflexes in dogs: rest and exercise. J Appl Physiol 95,
1159–1169.
Romer LM, Lovering AT, Haverkamp HC, Pegelow DF &
Dempsey JA (2006). Effect of inspiratory muscle work on
peripheral fatigue of locomotor muscles in healthy humans.
J Physiol 571, 425–439.
St Croix CM, Morgan BJ, Wetter TJ & Dempsey JA (2000).
Fatiguing inspiratory muscle work causes reflex sympathetic
activation in humans. J Physiol 529, 493–504.
Sheel AW, Derchak PA, Morgan BJ, Pegelow DF, Jacques AJ &
Dempsey JA (2001). Fatiguing inspiratory muscle work
causes reflex reduction in resting leg blood flow in humans.
J Physiol 537, 277–289.
Sheel AW, Derchak PA, Pegelow DF & Dempsey JA (2002).
Threshold effects of respiratory muscle work on limb
vascular resistance. Am J Physiol Heart Circ Physiol 282,
H1732–H1738.
Sinoway L, Shenberger J, Leaman G, Zelis R, Gray K, Baily R &
Leuenberger U (1996). Forearm training attenuates
sympathetic responses to prolonged rhythmic forearm
exercise. J Appl Physiol 81, 1778–1784.
Sinoway LI, Hill JM, Pickar JG & Kaufman MP (1993). Effects
of contraction and lactic acid on the discharge of group III
muscle afferents in cats. J Neurophysiol 69, 1053–1059.
Sinoway LI, Rea RF, Mosher TJ, Smith MB & Mark AL (1992).
Hydrogen ion concentration is not the sole determinant of
muscle metaboreceptor responses in humans. J Clin Invest
89, 1875–1884.
J Physiol 584.3
Sonetti DA, Wetter TJ, Pegelow DF & Dempsey JA (2001).
Effects of respiratory muscle training versus placebo on
endurance exercise performance. Respir Physiol 127,
185–199.
Spengler CM, Roos M, Laube SM & Boutellier U (1999).
Decreased exercise blood lactate concentrations after
respiratory endurance training in humans. Eur J Appl Physiol
Occup Physiol 79, 299–305.
Stuessi C, Spengler CM, Knopfli-Lenzin C, Markov G &
Boutellier U (2001). Respiratory muscle endurance training
in humans increases cycling endurance without affecting
blood gas concentrations. Eur J Appl Physiol 84, 582–586.
Tikunov B, Levine S & Mancini D (1997). Chronic congestive
heart failure elicits adaptations of endurance exercise in
diaphragmatic muscle. Circulation 95, 910–916.
Victor RG, Pryor SL, Secher NH & Mitchell JH (1989). Effects
of partial neuromuscular blockade on sympathetic nerve
responses to static exercise in humans. Circ Res 65, 468–476.
Volianitis S, McConnell AK, Koutedakis Y, McNaughton L,
Backx K & Jones DA (2001). Inspiratory muscle training
improves rowing performance. Med Sci Sports Exerc 33,
803–809.
Vrabas IS, Dodd SL, Powers SK, Hughes M, Coombes J,
Fletcher L, Demirel H & Reid MB (1999). Endurance
training reduces the rate of diaphragm fatigue in vitro. Med
Sci Sports Exerc 31, 1605–1612.
Wetter TJ, Harms CA, Nelson WB, Pegelow DF & Dempsey JA
(1999). Influence of respiratory muscle work on V̇O2 and leg
blood flow during submaximal exercise. J Appl Physiol 87,
643–651.
Acknowledgements
We thank our subjects for their enthusiastic participation. This
study was supported by the Natural Sciences and Engineering
Research Council of Canada (NSERC) and the Canadian
Foundation for Innovation. J.D.W. and J.A.G. were supported
by graduate scholarships from NSERC and the Michael
Smith Foundation for Health Research (MSFHR). A.W.S. was
supported by a Scholar Award from the MSFHR and a New
Investigator award from the Canadian Institutes of Health
Research.

C 2007 The Authors. Journal compilation 
C 2007 The Physiological Society